From: Kris Van Hees <kris.van.hees@oracle.com>
To: Alan Maguire <alan.maguire@oracle.com>
Cc: dtrace@lists.linux.dev, dtrace-devel@oss.oracle.com
Subject: Re: [PATCH v5 1/6] usdt: have copy_args() count args while parsing them
Date: Fri, 25 Jul 2025 11:57:24 -0400 [thread overview]
Message-ID: <aIOpZMcLIE14zBzs@oracle.com> (raw)
In-Reply-To: <20250710215930.82067-2-alan.maguire@oracle.com>
On Thu, Jul 10, 2025 at 10:59:25PM +0100, Alan Maguire wrote:
> stapsdt probes do not include an argument count, so the only
> way to count them is to parse the parameter string. Adjust
> copy_args() to set upp->sargc while parsing upp->sargv.
I am looking at reworking this a bit. The main challenge I see is that
it makes the overall copy_args() loop slower for *all* cases, not just
stapsdt. At a minimum, replacing the strlen(p) with a check for *p != '\0'
might be better.
But there is also the issue that given how late copy_args() is executed,
setting sargc become a moot point. It is not used after copy_args() is
called, so setting upp->sargc becomes pointless.
On the other hand, what we do want to know is how many arguments the probe
provides *before* we compile clauses, though in the adsence of real type
info, we could get away with not doing that. Having the number of args
allows the use of args[n].
> Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
> ---
> libdtrace/dt_prov_uprobe.c | 11 ++++++-----
> 1 file changed, 6 insertions(+), 5 deletions(-)
>
> diff --git a/libdtrace/dt_prov_uprobe.c b/libdtrace/dt_prov_uprobe.c
> index cf5cfd43..2cbd8910 100644
> --- a/libdtrace/dt_prov_uprobe.c
> +++ b/libdtrace/dt_prov_uprobe.c
> @@ -1145,9 +1145,9 @@ static void enable_usdt(dtrace_hdl_t *dtp, dt_probe_t *prp)
> }
>
> /*
> - * Generate code that populates the probe arguments.
> + * Generate code that populates, counts the probe arguments.
> */
> -static void copy_args(dt_pcb_t *pcb, const dt_uprobe_t *upp)
> +static void copy_args(dt_pcb_t *pcb, dt_uprobe_t *upp)
> {
> dtrace_hdl_t *dtp = pcb->pcb_hdl;
> dt_irlist_t *dlp = &pcb->pcb_ir;
> @@ -1158,7 +1158,7 @@ static void copy_args(dt_pcb_t *pcb, const dt_uprobe_t *upp)
>
> assert(pvp != NULL);
>
> - for (i = 0; i < upp->sargc; i++) {
> + for (i = 0; strlen(p) > 0; i++) {
> int ssize, disp, len;
> char *reg = NULL;
> int64_t val = 0;
> @@ -1425,6 +1425,7 @@ static void copy_args(dt_pcb_t *pcb, const dt_uprobe_t *upp)
> usdt_error(pcb, "Unknown format in arg%d spec", i);
> #endif
> }
> + upp->sargc = i;
> }
>
> /*
> @@ -1445,7 +1446,7 @@ static int trampoline(dt_pcb_t *pcb, uint_t exitlbl)
> dtrace_hdl_t *dtp = pcb->pcb_hdl;
> dt_irlist_t *dlp = &pcb->pcb_ir;
> const dt_probe_t *uprp = pcb->pcb_probe;
> - const dt_uprobe_t *upp = uprp->prv_data;
> + dt_uprobe_t *upp = uprp->prv_data;
> const list_probe_t *pop;
> uint_t lbl_exit = pcb->pcb_exitlbl;
> dt_ident_t *usdt_prids = dt_dlib_get_map(dtp, "usdt_prids");
> @@ -1519,7 +1520,7 @@ static int trampoline(dt_pcb_t *pcb, uint_t exitlbl)
> if (upp->flags & PP_IS_RETURN)
> goto out;
>
> - if (upp->sargc)
> + if (upp->sargv)
> copy_args(pcb, upp);
> else
> dt_cg_tramp_copy_args_from_regs(pcb, 0);
> --
> 2.43.5
>
next prev parent reply other threads:[~2025-07-25 15:57 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-10 21:59 [PATCH v5 0/6] add support for stapsdt probes Alan Maguire
2025-07-10 21:59 ` [PATCH v5 1/6] usdt: have copy_args() count args while parsing them Alan Maguire
2025-07-25 15:57 ` Kris Van Hees [this message]
2025-07-10 21:59 ` [PATCH v5 2/6] support stapsdt ELF-note-defined static probes Alan Maguire
2025-07-26 4:04 ` Kris Van Hees
2025-07-10 21:59 ` [PATCH v5 3/6] selftests/usdt: add test for stapsdt note-defined probe firing, args Alan Maguire
2025-07-10 21:59 ` [PATCH v5 4/6] selftests/usdt: add test for stapsdt notes in shared library Alan Maguire
2025-07-10 21:59 ` [PATCH v5 5/6] selftests/usdt: add test covering different forms of stapsdt note args Alan Maguire
2025-07-10 21:59 ` [PATCH v5 6/6] selftests/usdt: add test for stapsdt note-defined probe firing in -fPIE binary Alan Maguire
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=aIOpZMcLIE14zBzs@oracle.com \
--to=kris.van.hees@oracle.com \
--cc=alan.maguire@oracle.com \
--cc=dtrace-devel@oss.oracle.com \
--cc=dtrace@lists.linux.dev \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox